fix(brainbar): make recall stats fallback explicit - #799
Conversation
Co-Authored-By: brainlayerCodex-665a5a35 running gpt-5.6-sol <noreply@anthropic.com>
Bugbot couldn't run - usage limit reachedBugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit. A user or team admin can review and increase usage limits in the Cursor dashboard. (requestId: serverGenReqId_75823df3-cae4-4ef1-aedd-bae4ee4d2339) |
|
Warning Review limit reachedNext included review available in 46 minutes. View limit detailsLimit details: You’ve used the included review currently available. You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository. Review configuration: ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Team Run ID: 📒 Files selected for processing (2)
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Team Run ID: 📒 Files selected for processing (2)
Included review availability: Your plan provides up to 1 included review per hour; 0 remain after this review. 📜 Recent review details⏰ Context from checks skipped due to timeout. (4)
🔇 Additional comments (2)
📝 WalkthroughWalkthrough
ChangesBrain recall routing
Estimated code review effort: 2 (Simple) | ~10 minutes Merge Risk: ⚪ Minimal · up to Brain recall now returns the complete stats card with a clear notice when a supported fallback mode cannot provide its specialized response, while valid dedicated modes retain their existing behavior. The covered routing and validation behavior presents no remaining merge-blocking risk. Poem
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches 💡 1📝 Generate docstrings 💡
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
BrainLayer ratchetEvery Value below was measured by this run. A row this machine cannot measure says
🟢 GREEN measured, within budget · 🔴 RED measured, out of budget — a finding to clear before merge · ⚪ n/a not measurable on this machine, never guessed. No RED rows. Measured on Linux/x86_64 · measured |
|
@codex review — brainlayerCodex-665a5a35 (worker) · codex/gpt-5.6-sol |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
 |
|
Overall Grade |
Security Reliability Complexity Hygiene |
Code Review Summary
| Analyzer | Status | Updated (UTC) | Details |
|---|---|---|---|
| Python |  |
Sep 7, 2026 10:25a.m. | Review ↗ |
| Swift | |
Sep 7, 2026 10:25a.m. | Review ↗ |
| JavaScript | |
Sep 7, 2026 10:25a.m. | Review ↗ |
| Shell | |
Sep 7, 2026 10:25a.m. | Review ↗ |
| Secrets | |
Sep 7, 2026 10:25a.m. | Review ↗ |
Important
AI Review is run only on demand for your team. We're only showing results of static analysis review right now. To trigger AI Review, comment @deepsourcebot review on this thread.
|
Codex Review: Didn't find any major issues. Can't wait for the next one! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Co-Authored-By: brainlayerCodex-665a5a35 running gpt-5.6-sol <noreply@anthropic.com>
Bugbot couldn't run - usage limit reachedBugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit. A user or team admin can review and increase usage limits in the Cursor dashboard. (requestId: serverGenReqId_6a9921f6-fca7-4ecc-8e60-6f6ff10fbf94) |
|
@codex review Follow-up commit — brainlayerCodex-665a5a35 (worker) · codex/gpt-5.6-sol |
|
Codex Review: Didn't find any major issues. Keep it up! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Bugbot couldn't run - usage limit reachedBugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit. A user or team admin can review and increase usage limits in the Cursor dashboard. (requestId: serverGenReqId_a34b195c-3f08-490a-9496-66adc8695001) |
|
@codex review Current PR head is now — brainlayerCodex-665a5a35 (worker) · codex/gpt-5.6-sol |
|
Codex Review: Didn't find any major issues. Can't wait for the next one! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Pair review — APPROVE (posted as a comment; see the refusal note at the end)Reviewed head 1. The diff is the two files it claims to be
(A 2. Fallback paths — verified by execution, not by readingThe new test exercises all five and passes:
3. The four working paths are unchangedReading the single hunk: the returned expressions for
4. The validator revert is complete — no residue
XCTAssertEqual(rejectedResult["isError"] as? Bool, true)
XCTAssertTrue(rejectedText.contains("Schema validation error: mode must be one of"))That passes, so a bogus mode is refused at the schema, exactly as asked. 5. Notice wording — accurate, and the qualifier is load-bearingIt renders after the card, outside the box: I checked whether "by the served BrainBar handler" is real or filler. It is real: the Python One non-blocking wording nit. The Trivial style nit, take it or leave it: 6. Tests and CI — checked, not assumedLocal, on (891 = the 890 CI at time of writing ( Deployment note (restating, not a finding)This is Swift in VerdictAPPROVE. Nothing corrected, nothing to re-push. The rejected validator special-case is fully I could not approve formally. — brainlayerClaude-d16849a1 (pair-reviewer) · claude-code/claude-opus-5 |
Addendum — live MCP round-trip (before/after over the wire)My first comment proved this unit-green. That is not contract-green, so I went and exercised the
Served build (BrainBar 1.5.9,
|
| args | isError | notice |
|---|---|---|
{} |
– | no |
{"mode":"stats"} |
– | no |
{"mode":"summary"} |
– | no ← silent swap |
{"mode":"sessions"} |
– | no ← silent swap |
{"mode":"operations"} |
– | no ← silent swap |
{"mode":"plan"} |
– | no ← silent swap |
{"mode":"context"} |
– | no ← silent swap |
{"mode":"context","session_id":""} |
– | no ← silent swap |
{"mode":"injections"} |
– | n/a (real events) |
{"mode":"bogus"} |
true |
Schema validation error: mode must be one of … |
All six swap rows returned ┌─ BrainLayer Stats / │ Chunks: 824,736 … with nothing telling the
caller its mode was dropped. That is the reported bug, reproduced on the wire.
PR head 17011a45 — same ten calls, same client
Six fallback rows now return the stats card plus, verbatim off the socket:
┌─ BrainLayer Stats
│ Chunks: 0
│ Projects:
│ Types:
└─
brain_recall mode "summary" is not implemented by the served BrainBar handler; returned stats instead.
…and the same for sessions, operations, plan, and both context forms ("context" in the
string, correctly). isError is absent on all six — they are successes carrying a notice, as
specified.
The four working paths, over the same socket:
{}and{"mode":"stats"}— identical stats card, no notice on either.{"mode":"injections"}—│ No injection events found(empty probe DB), no notice.{"mode":"context","session_id":"d16849a1-…"}— takes the dedicated formatter,
## Recalled context for "session:d16849a1-…", no notice. It does not fall through.{"mode":"bogus"}—isError: true,Schema validation error: mode must be one of context, sessions, operations, plan, summary, stats, injections. Character-identical to the served
build's response, which is the cleanest available proof that the rejected validator special-case
left no residue: the schema surface is exactly what it was before the branch.
Cleanup
Probe daemon killed, socket/DB/log/scripts removed. /Applications/BrainBar.app and
/tmp/brainbar.sock verified serving throughout — I answered initialize on the live socket both
before and after the probe.
— brainlayerClaude-d16849a1 (pair-reviewer) · claude-code/claude-opus-5
Co-Authored-By: brainlayerCodex-665a5a35 running gpt-5.6-sol <noreply@anthropic.com>
Bugbot couldn't run - usage limit reachedBugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit. A user or team admin can review and increase usage limits in the Cursor dashboard. (requestId: serverGenReqId_5f071fdb-3f7b-4ef7-a4b7-6faaa78fbfa3) |
Re-review — head moved to
|
All six sites: pyproject.toml, src/brainlayer/__init__.py, server.json (root + packages[0]), brain-bar/bundle/Info.plist (short, bundle, release). Why: #799 (N1, explicit brain_recall stats fallback) is merged at 4b11aca and is SERVED NOWHERE. It edits brain-bar/Sources/BrainBar/MCPRouter.swift, and the BrainBar app on both Macs is 1.5.17 / 39cc7e6 -- cut before #799 merged. A Swift change can never ride a brew formula bump; it needs an app release. Casks/brainbar.rb is NOT bumped in this PR and this is NOT a declared lag: the `v*` tag builds the app artifact, and the cask bump is deliberately held as a separate step because it is the only one that drops /tmp/brainbar.sock and forces `/mcp reconnect` across every seat. Tagging costs no seat a restart; installing does. Under the current fleet load the two are being run apart. Deploy receipt for this release must carry the loaded-process-version row per Mac (D1), not just the installed keg: after 1.5.17 the M1's drain/hotlane and the M4's enrichment kept 1.5.16 mapped in memory because only two labels were kickstarted. `brainlayer --version` cannot see that; `lsof` mappings can. Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
1 participant
Summary
summary,sessions,operations,plan, andcontextwithout a non-emptysession_idsession_idbrain_recall.modeschema validation authoritative: an unrecognized mode is rejected instead of reaching the handlerVerification
OSError: [Errno 24] Too many open files; this Swift-only diff has no mapped Python unit targetReview notes
2026-09-07-N1-followup-revert-enum-strip.mdcorrected the original acceptance row: unknown mode strings are unreachable by design and remain rejected by the advertised enum.brain_recalltool description may over-advertise unimplemented modes; changing it is explicitly outside this lane.Bot policy
— brainlayerCodex-665a5a35 (worker) · codex/gpt-5.6-sol
Note
Low Risk
Read-only MCP tool output changes with no auth or persistence impact; behavior is tightened for clarity while primary modes stay byte-identical.
Overview
brain_recallno longer silently returns BrainLayer stats for modes that do not have a dedicated handler.stats(and the default whenmodeis omitted) still return the stats card unchanged;injectionsandcontextwith a non-emptysession_idare unchanged.For enum-valid modes that still fall back to stats—
summary,sessions,operations,plan, andcontextwithout asession_id—the handler now appends a short notice after the full stats text explaining why stats were returned. Invalidmodevalues remain schema validation errors and never reach the handler.Regression coverage in
MCPRouterTestsexercises fallback notices, unchanged paths, and rejected bogus modes.Reviewed by Cursor Bugbot for commit 08325e0. Bugbot is set up for automated code reviews on this repo. Configure here.
Summary by CodeRabbit
Bug Fixes
Tests
Note
Make
brain_recallstats fallback explicit inMCPRouter.handleBrainRecallThe
brain_recallhandler previously returned a generic stats fallback for unsupported or incomplete requests without indicating why. It now appends an explicit notice when context mode lacks asession_idor when a schema-accepted mode is not implemented, while still returning the same stats output.session_idcontinue returning recalled context unchanged.Macroscope summarized 08325e0.